www.xlr8yourmac.com

Mac Upgrades/Repairs | Storage | Video | Audio/Home Theater | OS Updates/Tips/Wi-Fi
Powermac G5 Hard Drive Heat/Cooling TipReturn to News Page
Powermac G5 Hard Drive Heat/Cooling Tip
(Relocating HD Temperature Sensor)

Posted: Dec. 2nd, 2003
Last Updated: March 16th, 2004
(Notes on 10.3.3 Update & Fan control Update for 10.2.7/10.2.8)


Apple G5 Fan Control Update for 10.2.7/10.2.8: Apple notes in the 10.3.3 update details that there's improved fan control for G5s. A reader sent a note they've also posted a Fan Control Update for G5 owners running 10.2.7/10.2.8:

"Not sure what's going on with this since I'm not sure if it's only 10.2.7 and 10.2.8 that are covered, but there appears to be a separate Fan Control Update 1.0 for PowerMac G5s. My guess is it's the same update found in the 10.3.3 update, but for G5 Jaguar users, but it does read "10.2.7 or later".
www.apple.com/support/downloads/fancontrolupdate.html (no longer online)
Take care. -Erik "

BTW - the reader that sent the article here last fall on relocating the G5's Hard Drive temperature sensor wrote that 10.3.3 seems to have addressed that (i.e. recurved the response for the sensor to better cool the drives.) His report is on the 10.3.3 feedback page and also mirrored below.

Comments on 10.3.3 Update Fan Control Improvements: (from the reader that sent the original sensor relocation info below)

(added 3/16/2004)
" This is the OS update I've been waiting for! Remember my submission to you a few months back concern the overheating of the G5 hard drives?
Well, with 10.3.3 we finally have evidence that Apple has recognized they did in fact have a hard drive heat problem in the G5. If you remember I moved my drive bay sensor from the roof of the case to the rear of my G5 to better detect drive bay heat and thus force the drive bay fan to spin faster and provide adequate cooling. Well, now it appears that modification is no longer necessary, as Apple has provided for much better cooling in that bay with 10.3.3.

At the time of my article, I predicted that Apple would probably produce a software update to reset the threshold temperature for the drive bay sensor in order to correct the overheating issue and avoid huge numbers of premature hard drive failures. Well, I can't say how many hard drive failed in the end, but they certainly took note of the problem, because 10.3.3. adjusts the threshold temp substantially as I predicted.

Here's the new data: with my sensor still in the new (hotter) location, the fan now spins much faster and much louder. Originally, the sensor in its original location would never reached the threshold temp, so the fan never kicked in above a slow twirl. In the new location, the fan works moderately to hold a target 104 degree temp.

Well, now the drive bay sensor has been reprogrammed for a much lower threshold to compensate for the poor original placement of the sensor. I've found that the new target threshold is about 87 degrees - 17 degrees lower than before! Clearly, Apple determined that the sensor in it's old location couldn't detect enough heat to cool the drives properly, so they lowered the threshold temp enough to force the drive bay fan to spin faster.

Of course, this means that my fan is now spinning TOO fast since the sensor is still in the new location. By leaving it in the new location, I will simply get even better cooling (the drive are actually cool to the touch), but the drive bay fan is actually very loud and annoying, so I will likely return the sensor to the original location. Hooray!

No doubt Apple probably saved a bunch of hard drives from premature failure in the coming hotter summer months.
Steve "


(original article text from Dec. 2003 follows)
The issue of hard drive heat in G5s has been reported several times in the past - some blaming the heat for drive failures (I'm not the only one to see drive failures with a G5, and at least one reader had repeated failures. Granted as a percentage of total users, it's probably small.) I can't prove the failures are due to heat, but cooler is always better. One common complaint has been that the hard drive cooling fan never seems to spin any faster than its initial low speed.
Last night a G5 owner sent a long email with comments, temperature tests and a tip on relocating the Hard Drive Cooling Fan Temperature Sensor to better cool the G5's hard drives. (copy of reader email follows)

"I have some interesting news that I think you readers may (eventually) find valuable. If you want to sit on this for a while, I would understand, but I wanted to get word to you now for when it might be useful.

I believe G5 owners (particularly those with 250GB drives, or especially 2 x 250GB drives) are very likely to experience hard drives failures within the first few months of use, depending on their environment. I have a lot of testing and personal experience to draw upon here (and even pictures, which are attached below). Here's the story:

About 2 weeks ago, my G5 (with 250GB Maxtor drive) began behaving strangely, including failing to clone properly to a firewire drive, kernel panics, and other unexplained behavior. I couldn't trace it to any single cause at first, because it would happen randomly. After intensive troubleshooting, including several fresh installs of the OS and many other ideas, I eventually determined that the problem was a failing hard drive. To test my theory, I even bought a second Maxtor 250GB and added it to the machine. When running off the new drive, everything was fine. When running the old drive, the problems returned. This was true even after I cloned the new drive onto the old drive. The old drive was simply failing mechanically. Toward the end of my testing, it even started to make grinding noises and wild head seeking sounds.
No big deal, right? Drives sometimes fail. Of course.

Well, here's the news: During the troubleshooting process, I began to notice that my drives were always very hot to the touch...extremely hot...burning hot. That bothered me. Even more troubling was the fact that whenever I looked at the drive bay fan, it was barely turning (I could almost count the fan blades as it turned). Clearly, the fan wasn't cooling properly and the drive likely failed due to excessive heat. That started me investigating the cooling system Apple was using in the G5.

Since then, I've done a lot of reading and poking around and experimenting, and I've believe I have good evidence to suggest that Apple has a serious problem brewing with the location of the drive bay temperature sensor. If I'm right, you're going to see a huge number of drive failures reported for this machine in the near future, and I predict you may even see Apple issue a recall to fix this problem, because it's going to continue happening even with replacement drives. In my case, Apple is sending me a new drive to replace my bad one.

I can I support these conclusions? I have lots of data. Let's start with the sensor itself. I've enclosed a picture of the sensor from my machine. Normally, it is affixed with glue directly to aluminum case at top of the drive bay above the drives. In my picture, it's shown hanging down because I've pulled it away from the glue (more on this later). It's about the size of a postage stamp.

G5 HD fan temp. sensor

While the top of the drive bay might seem like the right spot for the sensor (i.e. heat rises), as it turns out, it's not. There are two fundamental problems with Apple's placement. First, it's actually sensing the temperature of the air moving through the drive bay cavity, which is propelled by the drive bay cooling fan. Although the fan normally isn't moving the air very fast, it's still moving enough air to keep much of the heat from the drive from reaching the sensor. Of course, the air flowing across the sensor is substantially cooler than the drive itself (about 25 degrees cooler, according to my measurements).

Secondly, the aluminum case acts as a gigantic heat sink, sucking heat away from the sensor. So, even when the sensor begins to warm up from the heat of the drives below, it quickly dissipates its heat to the surrounding case, reducing the reading it reports. I was able to demonstrate both principles with a simple series of experiments. First, I used a program called ThermoGraphx to monitor the readings coming from each temperature sensor in the G5, including the drive bay sensor. ThermoGraphx accesses the same temperature data that the OS itself is monitoring to control the fans. To crosscheck this reading, I also measured the temperature of the drive bay and the hard drive itself using an infrared thermometer.

At the start of my test, the drive bay sensor read 98 degrees, yet my infrared measurement of the hard drive itself was 119 degrees. This high temperature is within Maxtor's specs, but it's still way too hot, in my opinion. You should also note the huge difference in temperature readings between the drive bay sensor and the hard drive itself. Then, I pointed a small external fan at that outside of the G5 case along the top rear portion of the machine (directly above the drive bay area). The fan blew room temperature air across the top of the case. Within minutes, the temperature reading from the drive bay sensor dropped nearly 10 degrees. I immediately opened the case and measured the temperature of the drive itself, which was actually a degree higher at 120! This suggests that people who use their G5 in a cold room (as I do) may be more likely to experience drive failure than those in a warm room, since the cold case fools the sensor into thinking the drive is cooler than it truly is!

Then, I turned my attention from the sensor to the fan. Why wasn't it turning faster? To see how much cooling potential the fan had, I removed the clear internal panel and replaced the outer aluminum door. Without the clear panel, drive fan will spin at max speed to protect the equipment. Within minutes of running with the fan on high, the drive bay temp reported near-room temperature readings, while the drive itself was only a few degrees above room temperature. So, obviously the fans could keep the drive cool, if only it would turn more.

My attention then turned to discovering what temperature was the fan programmed to hold? I assumed it must work like a air conditioner's thermostat. When the temperature rose above a certain threshold, the fan would begin to run faster to bring the temperature down. Since the fan never seemed to change speed (it always ran at the slowest speed), then I assumed it had never experienced a temperature high enough to exceed its programmed threshold. Maybe the threshold was set too low, especially given the location of the sensor. I devised a simple experiment to find out.

Using a halagon bulb from a desk lamp connected to lamp cord, I began exposing the sensor to heat while monitoring the fan speed and the temperature readings on ThermoGraphx. I removed all the hard drives from the G5 bay and booted the G5 from a firewire drive. I left the inner door on but removed the outer aluminum door. Before the test, the drive bay sensor read slightly above room temperature, reflecting the absence of any hard drives producing heat. Of course, the drive bay fan was barely spinning.

I began moving the light bulb closer to the sensor while monitoring the temperature reading on ThermoGraphX, Shareware Utility from www.kezer.net/thermographx.html. The reading moved up quickly, and as it passed about 107 degrees, the fan began to turn noticeably faster (faster than I had even seen before in normal use). After a few moments, I withdrew the light bulb a short distance and the temperature began dropping. Once it fell below 104, the fan began to slow down. With some experimentation, I found that fan speed began to increase between about 103 and 105 degrees and kept increasing as the temperature rose. I didn't test the upper limit for fear of damaging the sensor with the hot bulb (and as the fan blew stronger, it became more difficult to control the distance of the bulb from the element.

So it appears that the drive bay fan WILL NOT begin to blow a significant volume of air through the drive bay until the temperature at the sensor exceeds 104 degrees. Yet the temperature sensor is placed in such a way that it WILL NEVER read a temperature that high, even when the drive itself is nearly 120 degrees. Therefore, I believe most G5 hard drives are baking in their own heat due to little or no ventilation. Premature failures are inevitable.

Obviously, Apple should either reprogram their sensor threshold or move the sensor. I've remedied the situation for myself by moving the sensor (which is why I removed it from its mounting in the picture). I could only try a few locations given the limited amount of wire I had to work with. The best location I found is pressing the small circuit board directly against the hard drive housing along the right side of the drive (the side farthest from the drive fan). That location ensure the greatest transfer of heat from the drive to the sensor, plus the sensor is shielded from the direct flow of the air moving through the bay. Finally, it's "downwind" in the air stream, so it picks up the most heat possible. (see below for photos of the location)

Once I moved the sensor to this new location, my regular temperature readings jumped considerably. Similarly, my actual hard drive temperature measured with the infrared thermometer is only about a degree or two higher than the drive bay sensor reading. Finally, my drive bay readings are now sensitive to my use of the machine. Before I moved the sensor, the temperature would never move once it reached about 98-99 degrees. No matter how hard I worked the machine, the sensor rarely changed up or down.

Now when I work the drive hard (like when cloning the drive), the temperature will climb quickly to about 107-109, but immediately the drive bay fan will speed up and the temperature will drop back to about 105 for the remainder of the cloning process. When the drive is idle, the temperature will hold steady at 104. That confirmed my findings that the drive bay sensor is programmed for a threshold of about 104.

In conclusion, the G5's drive bay temperature sensor, when installed in its factory location, will never measure sufficient heat to activate the drive fan and provide sufficient cooling. Consequently, the drive is prone to overheating and failure. I'm giving you all this detail because you are the right person to monitor these reports. Interestingly, I have been browsing Apple's discussion boards, and I am reading a few reports of hot drives or even drive failures. Even more interestingly, I'm also seeing a LOT of reports of strange G5 problems similar to my early symptoms (unexplained kernel panics, random freezes, lost data, etc.). I can't help but wonder how many of those gremlins are the early signs of hard drive failure due to heat. We'll know if I'm right sooner or later, but I think this is a real problem which needs further attention.

Thanks for letting me dump all this on you. If you need more detail, please let me know.
Steve
(he later wrote)
The sensor is attached in the original location simply with a dab of silicon glue on the back side. It's actually easy to remove if you simple wiggle it loose. I just used the residual glue left behind on the board to affix it to the side of the hard drive. (you don't want any conductors/conductive parts of the sensor PWB contacting the hard drive case which could cause a short circuit.)

I will try to take a picture of the new location. It's bit difficult to see because it's hidden by the tracks of the drive mounts. I'll do my best, though, and I'll send that picture later today after I get home from work."

Steve later sent photos of the relocated sensor:

relocated temp. sensor

rear view of sensor

I asked Steve to comment about fan noise after the mod (and if any other case fans were running faster - which they should not be, since the relocated sensor is for the HD bay/fan only):

"I can tell you that the other fans are NOT spinning differently since I moved the drive bay sensor. Furthermore, ThermoGraphX reports the other 6 temperature sensors readings as no different since I made the mod than they before the mod, suggesting that the other sensors are distinct from the drive bay sensor (which makes sense, of course).

As far as sound levels, my G5 is not noticeably louder since I made this mod, despite the fact that the drive bay fan is spinning faster. Frankly, the drive bay fan is a rather quiet fan anyway, and in my case, the Radeon 9800 fan is the only fan I usually hear moving.
Overall, volume seems unchanged.
Steve"

My upper drive (Seagate 160GB, had failed previous) does feel a lot hotter after hours of use than the lower (Hitachi 160GB) drive. I'm going to try relocating the thermal sensor as soon as I can.

Steve's Comments after G5 Firmare Update 5.1.4: One of the most common questions about the G5 firmware update 5.1.4 (released Dec. 15th, 2003) was if it changed the hard drive bay temperature sensor curve/Fan response. (The HD bay sensor reports temps much lower than the actual drive in heavy use, and never turns the fans on high even when the upper drive is running hot - first noted earlier here on relocating the sensor.)
A reader that had moved the sensor previously ran tests again and reported the firmware update didn't address the original issue:

" I thought I should send you a quick follow up note concerning the G5 drive bay heat issue. After applying the recent G5 firmware update, I have done some brief testing. My drive bay sensor is still in its new location (against the side of the upper drive), and I did some testing to see if the firmware update improved the drive fan behavior.

From what I can tell at this point, the firmware update had no effect on the drive bay fan. The fan still starts off after a cold reboot running at the lowest possible speed and only begins to pick up speed after the drive bay fan senses a temperature higher than about 105 degrees. This is consistent with the behavior I experienced before the update.

Next, I temporarily reaffixed the sensor to the top of the drive bay. ThermoGraphX reported drive bay temperatures varying between the upper 80s to upper-90s over the course of several hours of light use. During that time, the fan never ran faster than the lowest speed and both drives got extremely hot to the touch. IR thermometer readings showed both drives running at temps exceeding 125 degrees by the end of the testing period (and some parts of the drives registered over 135 degrees!). This is exactly the kind of behavior I experienced before, which prompted me to move me drive bay sensor.

Based on this data, it seems fair to conclude that the firmware update was directed at altering drive bay fan behavior. Apple's own comments about the update would seem to support that conclusion, since they only mention correcting "open firmware" fan behavior.
Anyway, I have returned the sensor to the improvised location for now.
Steve "

Reader Feedback on Relocated Sensor: Here's the first feedback from a G5 owner that tried this tip (moved the sensor)

" After reading the Dec. 2 tip (above) on the location of the temperature sensor in the G5 drive bay, I thought I'd give my 2x2Gz with 2x250Gb drives a look. Sure enough, they were quite hot, although the drive bay temperature only read about 96F. Steve is right - the top of the bay and attached to the aluminum chassis probably isn't the best location if you want the sensor to reflect the actual drive temperature.

I then unstuck the sensor and attached it to the side of the drive that holds my system/home (figuring that's the one that gets the most use and will run the hottest - it was when I first checked them). (Note: Steve did not attach the sensor to a drive, but to the rear of the drive cage. As I said yesterday, don't allow any part of the sensor PWB/components to touch any metal surfaces to avoid the possibility of a short-circuit.) After it fired back up ThermographX read 113F (the highest by far I've ever seen), but that quickly dropped to 104F and I haven't seen it budge from there since - even after an hour of heavy disk use. After checking the drive temperatures by hand (the best I can do) they really are a lot cooler.

Cheers to Steve for that tip - he's likely saved my drives from an early death.

For any entrepreneurs out there - I'd imagine there's a market for these sensors on a longer length of wire - that way we wouldn't have to worry about voiding our warranties by un-gluing and moving the factory sensor. I just have to remember to move it back if I ever need service.
Again, thanks for the tip.
-Mike"

If drive heat is really the cause of the drive failures, moving the sensor may save Apple some warranty costs. (But of course their official policy on mods like this may invalidate the warranty, but at least this mod is easily undone and so far nothing but benefits seem to result.)

Other G5 Related Articles/Tips: See the G5 section of the Systems page for other articles on G5s Noise/tips, compatibility (PCI cards, etc.) and performance tests/comparisions.


= Popular Links =

macOS Updates


Mini-ITX Hackintosh


Dual M.2 NVMe
SSD PCIe Card
(Mac Pro 5,1)


Apple Guides


Display Tips


LG OLED TV Firmware


LG OLED TV Tests/Tips


Audio/HT Updates


TV, AVR, and Disc Player Tips


= back to www.XLR8YOURMAC.com =


= Other Site Topic Areas =
Mac Upgrades/Repairs | Storage | Video | Audio/Home Theater | OS Updates/Tips/Wi-Fi


Copyright © 2003. All Rights Reserved.
All brand or product names mentioned here are properties of their respective companies.

Legal: Site Privacy and terms/conditions of use.